Front. Phys.
DOI 10.1007/s11467-013-0305-0
REVIEW ARTICLE
Semiconductor nanowires for photovoltaic and
photoelectrochemical energy conversion
Neil P. Dasgupta1 , Peidong Yang1,2,†
1 Department
2 The
of Chemistry, University of California, Berkeley, CA 94720, USA
Center of Excellence for Advanced Materials Research (CEAMR), Chemistry Department,
King Abdulaziz University, Jeddah 21589, Saudi Arabia
E-mail: † p [email protected]
Received January 18, 2013; accepted February 12, 2013
Semiconductor nanowires (NW) possess several beneficial properties for efficient conversion of solar
energy into electricity and chemical energy. Due to their efficient absorption of light, short distances for minority carriers to travel, high surface-to-volume ratios, and the availability of scalable
synthesis methods, they provide a pathway to address the low cost-to-power requirements for widescale adaptation of solar energy conversion technologies. Here we highlight recent progress in our
group towards implementation of NW components as photovoltaic and photoelectrochemical energy
conversion devices. An emphasis is placed on the unique properties of these one-dimensional (1D)
structures, which enable the use of abundant, low-cost materials and improved energy conversion
efficiency compared to bulk devices.
Keywords nanowire, photovoltaics, artificial photosynthesis, photoelectrochemistry, solar energy
PACS numbers 62.23.Hj, 63.22.Gh, 84.60.Lw, 84.60.-h
Contents
1 Introduction
1
2 Nanowire photovoltaics
2
2.1 ZnO nanowire array PVs
3
2.2 Silicon nanowire array PVs
4
2.3 Single-nanowire PVs
5
3 Nanowires for photoelectrochemical energy
conversion
7
3.1 NW photocathode materials
8
3.2 NW photoanode materials
10
3.3 Integrated dual-bandgap NW PEC systems
11
4 Conclusion
12
Acknowledgements
11
References
13
1 Introduction
The dramatic increase in energy supply requirements facing humanity in the 21st century, combined with the negative environmental impacts of fossil fuel combustion,
have led to an urgent need for the large-scale deployment of renewable energy and energy efficiency technologies. Among the renewable sources of energy available on
earth, the solar flux represents by far the largest source,
with an average incident energy flux which is on the order
of 104 times larger than the total global power consumption. However, implementation of solar energy technologies has been limited by high costs and resource intermittency issues. Accordingly, there has been an explosion
in research on conversion and storage of solar energy,
with an emphasis on increasing efficiency and decreasing
manufacturing costs.
In order to efficiently convert solar energy to electricity or stored energy in chemical bonds, the solar resource
must be effectively absorbed, and this absorbed energy
must be extracted before losses due to heat dissipation
occur. Fundamentally, this requires three processes to
occur: absorption of a large portion of the electromagnetic radiation from the sun to generate excited charge
carriers, diffusion of charge carriers to an interface, and
separation of these excited charge carriers before recombination losses occur. Additionally, we must search for
c Higher Education Press and Springer-Verlag Berlin Heidelberg 2013
Neil P. Dasgupta and Peidong Yang, Front. Phys.
materials and device architectures which can facilitate
these processes without the need for expensive manufacturing techniques such as high vacuum or temperature
levels, or the use of precious, non-earth-abundant materials. These systems must be scalable to the terawatt
(TW) levels required for large-scale deployment of renewable energy on a global scale.
To address these challenges, nanostructured architectures have been explored due to a variety of advantages
they possess compared to “bulk” materials. For example, the characteristic length scales of solar energy conversion, such as the wavelength of light and the diffusion length of excited carriers in a semiconductor, are
typically on the nanometer to micrometer length scale.
This allows for novel phenomena such as optical and electrical confinement, stabilization of metastable material
phases, and geometric advantages such as high surfaceto-volume ratios, all of which can dramatically affect
the physics and chemistry of devices with components
at these length scales. Furthermore, there have been significant advances in fabrication techniques for nanomaterials in the past several decades, improving the ability to
reproducibly fabricate high-quality devices using robust
and high-throughput manufacturing methods.
Among the various categories of nanomaterials, semiconductor nanowires (NWs) are of particular interest in
solar energy conversion devices due to a combination of
several favorable properties [1, 2]. NWs can benefit the
absorption, diffusion, and charge separation steps of solar energy devices due to the unique properties of onedimensional (1D) materials at these length scales. For example, by fabricating vertical NW arrays the absorption
length (LA ) required to fully utilize the solar spectrum
can be decoupled from the carrier extraction length (LD ),
which should be minimized to reduce recombination [3],
as shown in Fig. 1. Additionally, NW templates suppress
reflection and enhance light scattering and trapping [4–
8]. This permits the use of significantly less material than
a planar architecture, and reduces the purity and morphological requirements of the absorber layer due to a
decreased carrier extraction length. As a result, NW devices have the potential to lower the cost of solar devices
both by decreasing the energy input required for material purification, and by enabling new material systems
based on abundant, low-cost elements which are not generally effective in planar geometries due to incompatible
absorption and diffusion length scales.
In this article, we will highlight recent advances by our
group in the application of semiconductor NWs to photovoltaic (PV) and photoelectrochemical (PEC) energy
conversion. This is not intended to be a comprehensive
review of the field, but rather a set of examples which
demonstrate the unique properties of these 1D materials.
LD
LA
2
Fig. 1 Vertical NW photovoltaic device, with a radial p-n junction. This allows for the photon absorption length (LA , associated
with the NW length) to be decoupled from the carrier diffusion
length (LD associated with the NW radius), as well as efficient
light scattering and trapping. This enables the use of abundant,
low-cost materials with reduced purity requirements and lower raw
material usage compared to bulk devices.
Our approach involves a combination of investigating
single-NW devices to gain a fundamental understanding
of the energy conversion and transport properties at the
nanoscale with the development of larger-scale NW array
devices for macroscopic energy conversion. An emphasis
will be placed on the unique characteristics and benefits
of using NWs for solar energy conversion, as well as the
associated challenges and directions for future research.
2 Nanowire photovoltaics
PV cells convert light energy directly into electricity
through excitation of electron-hole pairs in a semiconductor material, which are subsequently separated to external electrodes, providing a voltage and photocurrent
available for producing useful work. The current PV market is dominated by crystalline Si cells, which are able
to achieve efficiencies of around 25% [9], but are expensive to manufacture due to the high material costs. In
order to decrease the material cost, the use of cheaper
materials such as polycrystalline thin films or organic
semiconductors is possible. Additionally, managing light
in such a way that the solar resource is concentrated to
a smaller area or trapped inside the absorber layer of the
cell can decrease the amount of material required to fully
absorb the light. To further reduce the cost requires new
manufacturing techniques and architectures capable of
minimizing loss mechanisms such as recombination, reflection and parasitic resistances, while using low-cost
fabrication techniques and material systems. Due to the
aforementioned advantages of NWs, our group has been
exploring their implementation into vertical-array and
Neil P. Dasgupta and Peidong Yang, Front. Phys.
single-NW PV devices.
2.1 ZnO nanowire array PVs
Our initial efforts on NW PV array devices were based
on ZnO NW arrays given their favorable optical, electronic, and geometric properties. This was facilitated by
the development of low-temperature, solution-based processing of high quality single-crystalline ZnO NW arrays
[10, 11]. These arrays exhibit favorable properties for PV
energy conversion when combined with materials that
suffer from poor diffusion lengths such as polymers or
defective semiconductors. The solution-based processing
allows for a scalable, low-cost fabrication on a variety of
substrates, which could help drive down the costs of PV
manufacturing.
To demonstrate the ability to fabricate functional PV
devices using these ZnO NW arrays, a modified version
of the dye-sensitized solar cell (DSSC) was developed,
in which the mesoporous nanoparticle film used for dye
loading was replaced with a vertical ZnO NW array, as
shown in Fig. 2(a) [12]. This architecture presents several advantages compared to a nanoparticle film. While
the transport of charge carriers in the nanoparticle films
is dominated by trap-limited diffusion [13], the use of
vertical single-crystalline NWs provides a direct pathway for charge extraction, leading to faster charge collection in the NW cell. Furthermore, the NW geometry allows for the formation of a full depletion layer at
the semiconductor-electrolyte interface, leading to the
formation of a radial electric field which assists charge
separation. The complimentary electric fields in the radial and vertical directions assist charge separation and
extraction respectively, and allow for improved charge
transport compared to nanoparticle films. In addition,
an increase in the charge injection rate was measured
in the NW electrodes, due to a homogeneous crystallographic orientation along the NW surface.
The overall efficiency of the initial NW DSSCs was
limited to 1.5%, primarily due to the lower surface area
of the NW array compared to the nanoparticle film,
which limits light absorption. In Fig. 2(b), a comparison of Jsc vs. roughness factor is shown between the ZnO
NW electrodes, and nanoparticle films of TiO2 and ZnO.
The ZnO NW electrode shows similar behavior in the
TiO2 nanoparticle film despite being ∼10× longer in the
thickness direction, and better performance than both
of the ZnO nanoparticle films. This is direct evidence of
an improved charge collection due to the NW geometry.
A subsequent study demonstrated that the efficiency of
the ZnO NW DSSCs could be further improved by coating the NW surface with a thin TiO2 layer deposited by
atomic layer deposition (ALD), which led to an increase
3
in the Voc and fill factor due to a decrease in recombination at the semiconductor/electrolyte interface [14].
Further improvements in the efficiency would be possible
by increasing the NW surface area.
Fig. 2 (a) Schematic representation of a NW DSSC. (b) Comparison of Jsc vs. roughness factor between NW and nanoparticle
electrodes. The ZnO NW electrode exhibits superior performance
to ZnO nanoparticle electrodes, and exhibits similar performance
to the TiO2 electrode for roughness factors up to 200. Reproduced
c 2005 Nature Publishing Group.
from Ref. [12], Copyright Beyond DSSCs, ZnO NW arrays have been explored
for other material systems with short minority carrier
diffusion lengths in the absorbing layer. For example,
inorganic-organic hybrid NWs have been studied as a replacement for bulk heterojunction cells, due to improved
charge transport in the inorganic electrode with a direct current path, similar to the DSSC. It was shown
that an ALD TiO2 coating on hydrothermal ZnO NW
arrays leads to an improvement in performance in ZnOP3HT cells, due to an improvement in fill factor and
Voc [15]. This phenomenon was also demonstrated in alloxide NW PV devices, consisting of ZnO NWs in contact
with a p-type Cu2 O phase [16]. While the overall efficiencies of these devices remains below 1%, these studies demonstrated the importance of several critical factors, including the ability to form a high-quality interface
with low recombination rates to achieve improved performance in NW PV devices, while minimizing undesirable
losses due to shunting. If improved interfacial proper-
4
Neil P. Dasgupta and Peidong Yang, Front. Phys.
ties can be achieved, these architectures provide an ideal
platform to extract charge carriers from low-cost materials which suffer from short minority carrier diffusion
lengths.
2.2 Silicon nanowire array PVs
Due to the large elemental abundance and the vast
amount of research and development of Si processing
technology, it is likely that Si will maintain a significant portion of the PV market share in the near future.
Planar single-crystalline Si cells are a mature technology,
with only marginal room for improvements in efficiency.
Therefore, reducing the $/watt module cost will require
new ways to make inexpensive Si cells, while also maintaining high efficiencies.
To reduce the cost of Si PVs, our group has employed
two strategies. The first is to utilize significantly less raw
material, by improving the optical absorption of Si. Due
to the relatively low absorption coefficient of Si in visible wavelengths compared to direct bandgap semiconductors, a thickness of several hundreds of microns is
typically required to fully absorb the solar spectrum in
single-crystalline Si PVs. One way to mitigate this requirement is to increase the effective path length of light
in the material through optical scattering and light trapping strategies.
Several groups have investigated the anti-reflection
properties of microwire [17], nanowire [4, 6–8] and tapered nanocone [18, 19] geometries, which leads to a reduction in the losses at the top surface of the PV cell. In
addition to suppressing reflection, we have demonstrated
that nanowire arrays are very efficient at light trapping,
leading to an enhancement of the optical path length
compared to a planar absorber [5]. This permits the use
of significantly less material than a bulk crystalline PV
by reducing both the overall density of material in the
absorber layer, and the required thickness of the cell.
Light-trapping was investigated in top-down NW
arrays fabricated by nanosphere lithography and
anisotropic dry etching of crystalline Si substrates, as
shown in Fig. 3(a) and (b). This permitted simple and
accurate control of the NW diameter and length by controlling the size of the masking nanosphere and etching
time respectively. Radial p–n junctions were fabricated
by subsequent diffusion doping, and PV performance was
tested as a function of NW length. Interestingly, when
the NW length was increased from 8–20 µm, the shortcircuit photocurrent (Jsc ) only increased by 2%, compared to a 22% increase in planar control samples of the
same thicknesses. This suggested a strong light-trapping
effect in NW arrays, which was further investigated by
transmission measurements.
To quantify the light-trapping capability of the Si NW
arrays, free-standing Si-thin films on transparent SiO2
support layers were fabricated by back-etching silicon on
insulator (SOI) wafers, and transmission measurements
were performed through the thin membranes. These thin
Si films were subsequently patterned and etched to form
NWs, and the optical transmission was observed to decrease across a broad range of visible and near-IR wavelengths, as shown in Fig. 3(c). As the length of the NWs
increased, the transmission decreased, despite the fact
that the etching process was actually removing a significant amount material from the absorber layer. By
comparing the Jsc of thin PV devices with different NW
lengths, an increase in the optical path length enhancement was observed with increasing NW length. An increase in the path length of up to 73 was observed, which
exceeds the randomized scattering (Lambertian) limit
(2n2 ∼ 25) without a back reflector [20]. This demonstrates the ability of NW architectures to reduce the raw
material requirements to fully absorb the solar spectrum,
which could eventually lead to decreased PV manufacturing costs as our ability to fabricate uniform NW arrays on low-cost substrates continues to improve.
Fig. 3 Optical measurements of Silicon NW arrays of various
lengths. Cross-sectional SEM image of Si NW arrays with lengths
of (a) 2 µm and (b) 5 µm. (c) Transmission spectra of planar and
NW samples. NW samples exhibit significantly decreased transmission across a broad wavelength range due to increased lighttrapping in the NW arrays. Reproduced from Ref. [5], Copyright
c 2010 American Chemical Society.
Besides reducing the volume of raw material required
Neil P. Dasgupta and Peidong Yang, Front. Phys.
to fully absorb the solar irradiation, another strategy
to lower cost is to reduce the purity requirements of
the semiconductor material. For example, the cost of
metallurgical-grade Si is estimated to be on the order of
100 times lower than that of solar-grade Si [21]. However,
impurities in a semiconductor can lead to recombination
at defects, leading to a reduction in the minority carrier
diffusion length. This leads to an inherent mismatch between the material thickness required to fully absorb the
solar spectrum, and distance that excited carriers can
diffuse before recombining. As shown in Fig. 1, radial
p–n junctions in vertical NW arrays can mitigate this
problem, by orthogonalizing the absorption dimension
(associated with the NW length) and carrier extraction
dimension (associated with the NW diameter) [3].
To demonstrate this effect, we studied the performance
of Cu contaminated single-crystalline PVs [22]. To fabricate the devices, a thin film of Cu was sputtered on the
surface of p-Si wafers, and a high-temperature anneal
was performed to diffuse the Cu into the Si substrate.
This was followed by a wet etch to remove the Cu residue
remaining on the wafer surface. The Cu concentration in
the wafer was measured by Secondary Ion Mass Spectrometry, and a concentration of up to 1020 cm−3 was observed near the surface, which plateaus above 1017 cm−3
after the first micron of depth profiling. Subsequently,
nano- to micro-scale holes were etched to a depth of 6–10
µm into the wafer by deep reactive ion etching (DRIE).
Radial p–n junctions were formed by spin-coating and
diffusion annealing a phosphorous dopant, and finally,
metal top contacts were deposited.
The PV performance of these Cu-contaminated materials was tested by fabricating planar devices, which were
compared to Si devices without the impurities. A significant decrease in efficiency was observed in the “dirty”
cells, due to the presence of the impurities. However,
when the patterned radial p–n junction structure was
fabricated in the contaminated cells, an improvement in
the device efficiency was observed, with the greatest improvement measured for devices with a 4 µm pitch between the holes. The improvement in efficiency was due
to a larger short-circuit current in the radial devices,
which was attributed to improved collection efficiency
of minority carriers in the contaminated cells due to a
shorter diffusion length. For larger pitches, a smaller improvement in efficiency was observed. When the pitch
was reduced to 500 nm, a dramatic reduction in efficiency
was observed, to levels below the planar devices.
The dependence of efficiency on the periodicity of a radial p–n junction array can be understood by considering
the competition between improvements due to a shorter
minority carrier diffusion length, and the ability to fully
deplete the semiconductor to form a p–n junction. As
5
the pitch is decreased to values approaching LD , the
collection efficiency increases due to less recombination
from carriers that cannot effectively diffuse to the junction. Once the pitch is equal to or less than LD , a point
of “diminishing returns” occurs, where the volume over
which carriers are effectively collected begins to overlap,
no longer providing an improvement in short-circuit current. At this point, efficiency is likely to decrease, due to
an increase in surface recombination from a larger junction area. Eventually, when the pitch decreases below
the depletion width of the p–n junction, an overlap in
depletion regions occurs and the core-region is unable to
achieve a fully depleted diode, reducing the maximum
available Voc and efficiency.
These examples demonstrate several of the benefits
of NW PVs, including improved optical properties and
reduced material volume and purity requirements. Additionally, several challenges associated with balancing
efficient charge extraction, while minimizing surface recombination, have been identified. Further optimization
of NW PV structures will require a fundamental and
quantitative understanding of the effects of geometric
and material properties, and an improved control of the
interfaces. Accordingly, we have focused significant efforts on studying single-NW PV devices, which allow for
a careful study of the impact of these variables at the
nanoscale, while removing variables associated with naturally occurring variations in large ensembles of NWs in
an array and parasitic losses such as shunting pathways
resulting from macroscopic processing.
2.3 Single-nanowire PVs
Single-NW PV devices allow for an improved quantitative analysis of performance by reducing variations associated feature size, interfacial quality, and other nonuniformities in large-scale arrays [23–25]. For example, in
polymer-inorganic hybrid PVs, an improved understanding of the interfacial properties is crucial to increase device efficiencies [26], but can be difficult to analyze in
large-scale devices due to localized variations in material
properties and interface quality. To address this issue,
we fabricated individual NW inorganic-organic hybrid
PV devices consisting of a single-crystalline ZnO NW
core and an organic shell of end-functionalized oligo- and
polythiophenes, which were grafted onto the NW surface
using solution processing techniques to form a radial pn junction [27]. Individual NWs were characterized by
TEM, showing a core-shell structure, allowing for a visualization of the morphology of the inorganic-organic interface. Planar single-NW devices were fabricated on an
insulating substrate, and individual contacts were made
to the core and shell regions. The J–V characteristics
6
Neil P. Dasgupta and Peidong Yang, Front. Phys.
showed higher Voc values than array devices, suggesting that the quality of the grafted interface is higher
than that formed by solution processing. On the other
hand, the devices had lower Jsc values than theoretically
predicted, illustrating the limitations of absorption and
charge separation in these organic absorber materials.
Despite the relatively low overall efficiencies of these devices, the single-NW PV platform proved to be an effective platform to study the various factors affecting device
performance in macroscopic cells.
While NW-based devices have demonstrated various
optical and charge separation benefits, they are still typically less efficient than planar devices of the same material system [24, 25, 28]. This is typically related to
a lower Voc and fill factor in the NW devices due to increased interfacial recombination from the large junction
surface area. This is particularly evident in the inorganicpolymer hybrid devices. To address this issue, we developed a simple method to produce high-quality epitaxial
interfaces in radial p-n junctions by using mild solution
processing [29]. A cation exchange reaction was used to
form epitaxial CdS/Cu2 S core-shell NWs, which exhibited Voc and fill factors of 0.6 V and 80% in a single NW
device (Fig. 4). These values are larger than equivalent
planar devices, with a lower drop in Voc in the NW devices with reduced light intensity. This is attributed to
the high-quality interface in this single-crystalline coreshell structure, leading to improved collection efficiency.
An overall efficiency of 5.6% was measured, which was
limited only by a lower Jsc than the theoretical maximum, due to insufficient absorption of visible wavelengths in the thin Cu2 S shell. This suggests that increas-
ing the shell thickness, while maintaining the benefits of
the high-quality interface would lead to even higher efficiencies.
An additional benefit of fabricating horizontal devices
is the ability to study the uniformity of charge collection
as a function of position in and individual NW. To accomplish this, scanning photocurrent mapping (SPCM)
was performed, demonstrating that the core-shell region
was active for charge separation, with a sharp drop-off at
the edge of this region. This illustrates the importance of
a radial p-n junction in this system, which benefits collection of charge from the CdS material which suffers from
a low minority carrier diffusion length (<1 µm). Furthermore, multiple devices could be fabricated from a single
NW, allowing for series and parallel configurations which
are current or voltage matched due to the use of the same
NW. This demonstration of an improved performance in
a NW PV device compared to planar equivalents using
simple, and potentially low-cost processing, is an important indication of the potential benefits of core-shell NWs
in solar cells.
Beyond studying the importance of the radial p-n interface quality on device performance, single-NW devices
can be used to quantify the optical benefits associated
with subwavelength interactions between light and the
nanostructure. For example, individual NWs have been
shown to support leaky-mode resonances, which can
enhance light absorption at resonant wavelengths [30,
31]. Additionally, metal nanoparticles are being actively
studied as a way to enhance absorption in PV materials
through plasmonic manipulation of light [32–34]. Both
of these phenomena involve interaction of light with
Fig. 4 Core-shell CdS/Cu2 S single-NW PVs. (a) Schematic of the fabrication process. A core-shell structure is formed by solutionbased cation exchange of the CdS NW in a solution containing Cu ions. to form a Cu2 S shell. (b) SEM image of a single-NW PV
device. (c) I–V curve of a single-NW PV device under 1 sun (AM 1.5 G) irradiation. (d) Variation of Isc and Voc on light-intensity
c 2011
(e) Wavelength-dependence of photocurrent compared with simulated absorption. Reproduced from Ref. [29], Copyright Nature Publishing Group.
7
Neil P. Dasgupta and Peidong Yang, Front. Phys.
subwavelength features, which could potentially increase
the performance of PV devices.
To study these effects, we fabricated a single-NW Si
PV device decorated with an octahedral silver nanocrystal [35]. Si NWs were grown in a bridging configuration between two electrodes, using VLS growth in a
CVD reaction. Radial p-n junctions were fabricated using a series of lithography and masking steps followed
by vapor-phase doping. To study the absorption properties of these NWs, wavelength-dependent photocurrent
measurements were taken on the devices. These measurements show that the NW devices exhibit peaks in photocurrent associated with optical resonances of the wire,
which were in agreement with numerical simulations of
the absorption using a three-dimensional finite-difference
time domain (FDTD) method.
When a single silver nanocrystal was deposited onto
a single-NW PV device, changes in the wavelengthdependent photocurrent were observed. The experimental results showed an increase in photocurrent at
wavelengths associated with plasmon resonances in the
nanoparticle, and a decrease at wavelengths associated
with the optical resonances of the NW (Fig. 5). These
effects are in agreement with FDTD simulations of the
NW-nanocrystal system, which predict a perturbation
of the NW resonances by the nanoparticle. This study
demonstrates the power of single-NW PV devices to
quantitatively study the optical properties of nanostructured geometries, providing insight into the design of fu(a)
ture devices which could benefit from enhanced absorption from these subwavelength effects.
Our research in NW PV devices has demonstrated
many benefits of these structures associated with improved light absorption and charge separation in radial
p-n junctions. The importance of a high-quality interface has been demonstrated, and a quantitative analysis of individual interfaces has been achieved by use of
single-NW devices. The importance of balancing these
benefits with potential disadvantages associated with increased surface/interface recombination remains a critical engineering challenge, which requires further insights
into the effects of geometric and material variables on device performance. Nevertheless, significant progress has
been made towards the development of efficient NW PVs,
which could dramatically reduce the cost of PV technology by decreasing material loading and purity requirements, while maintaining high efficiencies. In the next
section, we will discuss how many of these same benefits can be utilized to design photoelectrochemical (PEC)
cells, which store solar energy in the form of chemical
bonds, which can later be utilized as a renewable fuel
source.
3 Nanowires for photoelectrochemical energy
conversion
Due to the intermittency of the solar resource, energy
(b)
nA
4.0
p contact
p contact
3.5
3.0
2.5
2.0
1.5
1.0
n contact
Fractional change in absorption
(c)
0.5
n contact
0.0
(d)
0.3
0.2
100 nm
0.1
0
0.1
400
450
500
Wavelength /nm
550
600
Fig. 5 (a) SPCM of a single-NW Si PV decorated with an octahedral silver nanocrystal. (b) Corresponding scanning
optical image collected simultaneously showing scattering at the position of the nanocrystal. The scans were taken at a
wavelength of λ = 442 nm. (c) The enhancement at the location of the nanoparticle is in agreement with a FDTD simulation
c 2011 American Chemical Society.
for a NW this size. (d) SEM images of the device. Reproduced from Ref. [35], Copyright 8
Neil P. Dasgupta and Peidong Yang, Front. Phys.
storage is a critical component to a renewable and sustainable energy future based on using sunlight as an energy source. Furthermore, a large component of our overall energy use is in the form of transportation fuels, which
require a means to store the diffuse solar resource in a
more concentrated resource. To address these issues, we
have been investigating the direct conversion of solar energy into chemical fuels, through a process known as artificial photosynthesis [36–38]. In this process, we attempt
to mimic what nature does every day in photosynthesis
by storing the solar energy resource in the form of chemical bonds. The inputs to this process include sunlight,
CO2 and water, and the outputs could be in the form
of gaseous or liquid fuels such as hydrogen, methanol,
or longer-chain hydrocarbons. Besides storing solar energy for later use, this process attempts to close the cycle
of CO2 emissions associated with combustion reactions,
resulting in a truly carbon-neutral process.
The implementation of artificial photosynthesis entails
several strict requirements. In addition to the absorption
and charge separation requirements of PVs, we must utilize excited charge carriers to drive electrochemical reactions in an efficient manner. This is typically done by
forming a semiconductor-electrolyte interface, which allows for charge transfer reactions to occur driving the
redox reactions in solution [39]. However, the kinetics of
these reactions must occur at a high enough rate to extract the excited carriers before recombination occurs,
which often time requires the inclusion of a co-catalyst
material at the interface and loss of energy to electrochemical overpotentials. Furthermore, competing reactions such as corrosion or redox chemistry involving the
semiconductor material can detrimentally affect the performance and lifetime of the devices. Therefore, we must
seek robust and stable materials which can withstand the
harsh conditions of PEC reactions in both a reducing and
oxidizing environment.
While some semiconductor materials such as TiO2 for
solar water splitting have been shown to address many
of these challenges [40], high efficiencies for single materials have been limited by insufficient absorption of the
visible spectrum from high-bandgap metal oxide materials. To address this problem, researchers have shifted
their focus to a PEC diode structure based on a coupled
photocathode and anode material [41–43]. By splitting
the overall reaction into two half-reactions at different
interfaces,one can absorb a larger portion of the solar
spectrum by a two-photon absorption process, which is
similar to natural photosynthesis.
Semiconductor NWs possess several advantages for the
implementation of artificial photosynthesis schemes [38].
Many of the favorable optical and electronic properties
discussed in the NW PV section apply to photoelectrode
materials, including improved light absorption, a minimized minority carrier diffusion length, and high-quality
single crystalline materials. Additionally, the large surface areas associated with NW arrays can reduce activation overpotentials, by reducing the kinetic losses compared to planar devices. Therefore, we have focused our
efforts on developing individual photocathode and anode
materials, as well as developing strategies for integrating
these materials into a dual-bandgap device capable of
performing artificial photosynthesis. In the following sections, we will highlight recent examples of work on these
individual components, and progress towards a fully integrated dual-electrode device.
3.1 NW photocathode materials
The photocathode is the location of the reduction halfreaction, where the fuel source is generated. The simplest photocathode reaction is the reduction of protons
in water to produce hydrogen. This reaction involves the
transfer of two electrons from the semiconductor to the
electrolyte, producing one hydrogen molecule. Several
candidate materials which have conduction band edges
higher in energy than the H+ /H redox potential with
suitable bandgaps for visible light absorption have been
identified. Our research has focused on p-Si and p-GaP
due to their favorable bandgaps and sufficient photovoltages to contribute substantially to the 1.23 V required
for overall water splitting in a dual-bandgap device.
Silicon is a popular photocathode material due to its
elemental abundance, favorable band alignment for the
hydrogen evolution reaction, and the vast amount of
knowledge in Si processing for PV materials [44, 45]. We
have developed several methods for fabricating Si NW arrays, as highlighted in the NW PV section, based on both
bottom-up and top-town processing. Due to relatively
sluggish kinetics of hydrogen evolution, a co-catalyst is
typically added to decrease the activation overpotential,
which is typically Pt [39, 46] or MoSx [47, 48]. These Si
photocathodes have shown the ability to produce high
photocurrent densities of over 20 mA/cm2 , which are required to achieve high conversion efficiencies in tandem
devices. Our research is currently focused on reducing the
cost of these materials through similar strategies as the
Si NW PV array devices, by reducing material volume
and purity requirements. Additionally, we are seeking to
reduce the cost associated with the catalyst by searching
for earth-abundant materials and seeking to reduce the
Pt loading requirements by nanostructuring.
An alternate photocathode material of interest is GaP
[41, 49], due to a conduction band edge of ∼1 V more
negative than both the water reduction and CO2 reduction potentials and relatively small bandgap (2.23
Neil P. Dasgupta and Peidong Yang, Front. Phys.
eV at 300 K). However, the majority of the GaP NW
growth techniques reported have used high-temperature
vapor phase approaches, which could add significantly to
manufacturing costs of large-scale systems. Therefore, we
developed a low-temperature solution-liquid-solid (SLS)
growth process, which is capable of generating large scale
quantities of NW product [50]. A key aspect of this
process is that unlike previous SLS NW growth processes, the GaP NWs can be grown without the presence
of surfactant molecules on the surfaces (Fig. 6). These
surface species can hinder charge transfer reactions at
the semiconductor-electrolyte interface, and are therefore not desirable for artificial photosynthesis.
The SLS process is similar to VLS NW growth, in
which a metal catalyst particle becomes supersaturated
with a reactant element, and a single-crystalline NW is
precipitated which grows in a 1D manner. However, in
this case, the reaction occurs in solution, allowing for
a highly scalable reaction, which is not dependent on
substrate surface area. Our GaP synthesis consists of injection of Ga and P precursors in a 1:1 molar ratio into
a hot squalane solvent. The Ga precursor thermally decomposes to form nanoscale Ga metal droplets, which
subsequently act as catalysts to promote the growth of
GaP NWs. The resulting product consisted of GaP NWs
with lengths of 1–2 µm, and diameters of ∼30 nm. Finally, the wires were cleaned and the Ga nanoparticles
were selectively etched in an HCl solution, leaving the
wires undamaged. A structural analysis by HRTEM and
XRD confirmed that the wires were single-crystalline
GaP, with visible light absorption measured by uv-vis
spectrometry.
The PEC performance of the wires was measured by
drop-casting a suspension of the wires in acetone onto
a metal-coated glass substrate, followed by an inert gas
9
anneal to make ohmic contact. The wires exhibited a
positive increase in Voc of ∼0.15 V under AM 1.5 1 sun
illumination, indicating photocathodic behavior despite
the fact that the wires were not intentionally doped. The
ability to generate H2 gas under visible-light illumination
in neutral pH was measured using gas chromatography
(GC) using methanol as a hole scavenger, and the H2
evolution rate was increased by ∼1 order of magnitude
when small Pt catalysts were loaded on the NW surface.
Additionally, the stability of the wires under illumination
in solution was confirmed by HRTEM images before and
after a 18 h GC measurement, showing no noticeable
difference in the NW surfaces.
In addition, we have developed a way to intentionally
dope these wires with Zn by inserting diethylzinc (DEZ)
into the reaction solution [51]. The Zn concentration in
the wires could be controlled by varying the amount of
DEZ injected into the solution. The p-type behavior of
these wires was confirmed by single-NW field effect transistor (FET) measurements. The Zn-doping resulted in
an increase in open-circuit photovoltage (up to 280 mV),
and Mott–Shottky measurements gave a flatband potential of 0.52 V vs. RHE. An increase in carrier concentration due to the dopants results in an improved photovoltage and photocurrent compared to undoped wires, which
is attributed to a reduced depletion region width to below the NW radius, preventing complete depletion of the
NW. This is a central issue to both NW PV and PEC
devices, indicating the importance of doping control in
nanostructures.
An optimal doping concentration of 0.6%–1% Zn was
found, leading to a photocurrent which is ∼80% that
of planar GaP. Despite this lower value, the absorbed
photocurrent efficiency (APCE) of the NW sample was
consistently higher across visible wavelengths than the
Fig. 6 Large-scale, solution synthesis of semiconductor nanowires. (a) As-synthesized GaP nanowires in 1 L of hexadecane. (b)
A GaP nanowire membrane fabricated by filtration of water-dispersed GaP nanowires through a filter paper (white). After drying,
the nanowire membrane became free-standing. (c) Transmission electron microscopy image of the GaP nanowires. Reproduced
c 2011 American Chemical Society.
from Ref. [50], Copyright 10
Neil P. Dasgupta and Peidong Yang, Front. Phys.
planar sample, showing improved charge separation efficiency. Furthermore, the APCE was shown to increase
with decreasing loading of the NWs suggesting that
charge transport across the NW/NW contacts limits the
overall performance. Thus, with improved contact between the wires, the photocurrent could potentially be
improved further. This study demonstrated the advantages of the NW geometry for PEC performance in GaP,
by achieving comparable results while utilizing 1/3000
of the quantity of GaP in a planar wafer, with a higher
APCE over the visible spectrum.
3.2 NW photoanode materials
The photoanode presents a more challenging set of requirements, due the photoanodic corrosion under of
many semiconductor materials with appropriate band
edges for the water oxidation reaction [39]. Therefore,
metal oxide materials have been widely studied due to
their stability under photoanodic conditions. However,
due to the relatively low valence band edges of metal oxides compared to the water oxidation potential (+1.23 V
vs. NHE), a significant loss in energy results due to the
valence band offset. Furthermore, many metal oxide materials have large bandgaps, which limit the absorption of
visible light, and therefore the overall photocurrent. Thus
the search for an efficient photoanode material typically
involves three approaches: increasing the visible light absorption of wide-gap metal oxides, thin surface coatings
of unstable materials to improve stability against corrosion, and searching for new stable materials which absorb
a greater portion of the visible spectrum.
Since the demonstration of water splitting using TiO2
by Fujishima and Honda in 1972 [40], TiO2 has been
widely studied as a photoanode material due to its relative stability and low cost. However, due to the high
bandgap of rutile TiO2 , as well as a low electron mobility
and minority carrier diffusion length, the resulting photocurrents are too low for efficient implementation into
a dual-bandgap tandem PEC device. NWs have the potential to improve the quantum efficiency of TiO2 in the
UV, by decoupling the absorption and diffusion lengths,
as shown in Fig. 1. However, an increase in length also
has certain disadvantages, such as increased surface recombination and difficulty in electron extraction due to
a poor electron mobility.
To improve our understanding of these effects, we performed a study on rutile TiO2 NW arrays to quantify the
dependence of PEC performance on NW length and surface coatings [52]. Rutile TiO2 NW arrays were grown
on FTO substrates by a hydrothermal method, which
allowed for control of NW length by varying the growth
time. PEC measurements were performed on 4 differ-
ent lengths of NW arrays, and an increase in the anodic photocurrent with increasing NW length from 0.28
to 1.28 µm was observed, which appeared to approach
a saturation current. Furthermore, the IPCE spectra
showed a red shift in the maximum EQE with increasing NW length. This suggests that carrier extraction
can limit performance in longer wires for shorter wavelengths, which are absorbed near the top of the wires.
To further improve the energy conversion efficiency
of TiO2 NWs, a thin surface coating of TiO2 was deposited on the NW arrays by ALD. This strategy has
been used to improve charge collection efficiency by reducing surface states [53]. The ALD films were grown
at 300◦ C using TiCl4 and water as precursors. HRTEM
and XRD analysis showed a variation in the phase of the
ALD TiO2 material from partially amorphous, to rutile,
to anatase with increasing shell thickness. Interestingly,
the rutile phase formed epitaxially on the NW core, as
confirmed by HRTEM. The PEC performance of these
coatings demonstrated a decrease in anodic photocurrent
for the amorphous and anatase coatings, and an increase
with the rutile coating. The increase in photocurrent for
the rutile coating was attributed to a combination of increased surface passivation, increased surface area due to
the polycrystalline shell, and an improvement of charge
collection efficiency, as shown by an increase in the IPCE
for shorter wavelengths.
Despite these improvements in TiO2 NWs, the overall
energy conversion efficiency of TiO2 is limited by low absorption due to a high bandgap. Therefore, many groups
are studying the ability to use TiO2 or other stable oxide materials as surface coatings to protect otherwise unstable photoanode materials with smaller bandgaps [54].
This strategy requires the use of a thin and conformal
coating layer, capable of oxidizing water at the interface between the coating layer and electrolyte, while preventing corrosion of the underlying light absorber under
highly oxidizing conditions in solution. To demonstrate
this effect, we deposited thin layers of TiO2 on planar
and NW Si substrates using ALD [55]. In this structure,
UV-light is absorbed in the TiO2 , while visible light is
absorbed in the Si. Holes generated in the TiO2 oxidize
water, while electrons generated in TiO2 recombine with
holes generated from the Si at the Si/TiO2 interface.
For both n-type and p-type Si, an increase in the saturation photocurrent of 2.5× that of planar substrates
is observed in the NW samples, which is attributed to
decreased reflectance and increased surface areas. Additionally, n-Si/n-TiO2 samples exhibited higher photovoltages and photocurrents than p-Si/n-TiO2 samples,
due to favorable band-bending at the semiconductor heterojunction interface, which reduces losses from holes recombining at the heterojunction interface due to favor-
Neil P. Dasgupta and Peidong Yang, Front. Phys.
able band bending. Thus, while the oxidation photocurrent still comes from absorption in the shell material, the
n-n junction can improve the short circuit current by decreasing recombination at the back surface of the shell,
while providing a shift in VOC towards more negative
onset potentials.
While stabilizing lower-bandgap semiconductors by
surface coating addresses the absorption and stability criteria separately, the ideal photoanode should have both
of these qualities in one single material. Although several
metal oxides have demonstrated stability against photoanodic corrosion, the valence band maximum is typically more than 1 eV below the redox potential for water oxidation. Therefore, much of the energy absorbed
in oxide materials does not contribute to the useful energy extracted through water splitting. An alternate set
of compounds is the metal nitride family, which has a
higher valence band maximum due to the smaller ionization energy of N 2p orbitals compared to O 2p orbitals.
In particular, Inx Ga1−x N alloys have been shown to have
a tunable bandgap from the UV to near-IR region of the
spectrum [56]. This allows for tuning of the absorption
properties to allow for a material capable of either overall water splitting, or acting as an efficient photoanode
material.
We have previously developed a halide chemical vapor deposition (HCVD) technique for growing singlephase InGaN wires width tunable stoichiometries [56].
To study the PEC properties of these wires as a photoanode, Inx Ga1−x N (x = 0.08−0.1) were grown on planar Si
and Si NW array substrates (Fig. 7) [57]. The InGaN/Si
hierarchical arrays provided a photocurrent at an onset
voltage of 0.1 V vs. RHE, and reached a photocurrent
density of 33 µA/cm2 at 1.23 V vs. RHE. This was 5
times higher than the NW arrays grown on planar Si substrates, which attributed to a higher surface area for the
electrochemical reaction and increased optical pathway
in the hierarchical arrays. Furthermore, these NWs were
11
stable against photoanodic corrosion or oxidation after
15h of continuous illumination in a pH=3 electrolyte at
a high light intensity (350 mW/cm2 ), as confirmed by
HRTEM analysis.
Despite their promising onset voltage and stability, the
photocurrent for these devices is only 1.2 % of the theoretical maximum based on their bandgap. This low current was attributed to fast recombination and sluggish
charge transfer kinetics at the semiconductor interface.
Efforts to improve the material quality and addition of a
co-catalyst are potential areas of improvement for these
structures. Therefore, the search for an efficient photoanode material continues, and is an important direction for
future research.
3.3 Integrated dual-bandgap NW PEC systems
The final goal of the dual-bandgap PEC research is to
generate a system capable of performing water splitting
and artificial photosynthesis using only solar energy as
an input. To achieve this goal, we need to develop architectures capable of integrating a photoanode and photocathode into a single, current matched device. To accomplish this difficult task, we have proposed two approaches: an asymmetric junction based on a core-shell
structure with both materials present in a single NW,
and a bilayer NW fabric.
As model system for demonstrating a two-bandgap
scheme, we have developed the ability to fabricate a
Si/TiO2 asymmetric junction onto single NWs [58]. In
this case, the Si acts as the photocathode where the reduction reaction occurs, and the TiO2 acts as a photoanode to oxidize water. Asymmetric core-shell NW structures were fabricated by growing Si NWs via a VLS
mechanism, which were subsequently coated by TiO2
shell layers by ALD. These shells were removed from a
part of the NW by masking and selective etching of the
Fig. 7 High surface area Si/InGaN nanowire photoanodes. SEM images of hierarchical Si/Inx Ga1−x N nanowire arrays on
a Si (111) substrate with x = 0.08–0.1 (a). A fractured wire reveals the cross-section (b) showing that InGaN nanowires
c 2012 American Chemical Society.
grow vertically from the six Si wire facets. Reproduced from Ref. [57], Copyright 12
Neil P. Dasgupta and Peidong Yang, Front. Phys.
Fig. 8 Schematic illustration (left) and a zoomed-in view (right) of a bilayer nanowire fabric-based composite membrane
for direct solar water-splitting. The top layer is a high-surface-area nanowire mesh decorated with oxidation catalysts.
This part serves as a photoanode responsible for water oxidation; the bottom layer is a high-surface-area nanowire mesh
decorated with reduction catalysts. This is the photocathode responsible for water reduction. Reproduced from Ref. [59],
c
Copyright 2012
Nature Publishing Group.
TiO2 , leaving an exposed Si surface.
A key necessity for the successful implementation of
a dual-bandgap PEC system is the efficient recombination of majority carriers between the two materials,
while the minority carriers drive the separate redox reactions at the anode and cathode. The asymmetric NW
structure allows for a study of this charge separation
phenomenon by probing a single NW. To demonstrate
this effect, asymmetric Si/TiO2 NWs were illuminated
by UV light, and the surface potential was measured as
a function of position on the NW using Kelvin probe
microscopy (KPM). This is a scanning probe technique,
in which a conductive atomic force microscope (AFM)
tip is used to measure surface potential with nm resolution. Controlled humidity was used to condense water on
the NW surface, leading to the formation of a semiconductor/electrolyte interface for both materials. An increase in the positive surface potential was measured in
the TiO2 region relative to the Si region upon illumination, which is attributed to a buildup of holes at the TiO2
electrolyte interface, and electrons at the Si/electrolyte
interface. This demonstrates the ability of asymmetric
structures to separate minority carriers while allowing
majority carriers to recombine at the interface.
While the asymmetric structure provides a model system for studying a dual-bandgap system at the single
NW system, the fabrication of such structures requires
expensive processing steps, and the separation of the
product species will not be trivial. Therefore, we have begun to develop bilayer NW fabrics, based on solution processing and flexible, low-cost substrates [59]. The photoelectrode materials will consist of NW meshes with high
surface area, using solution-processed NW materials such
as our SLS grown NWs (Fig. 8). A proton conducting material will be placed between the electrode materials, and
matching of the current densities will be required for effi-
cient utilization of the solar resource. This will reduce the
costs associated with high-temperature or vacuum processing, separation of product species, and the turnover
frequency requirements of co-catalysts due to the use of
high surface area electrodes. The major challenge now
is to develop efficient photoanode materials, capable of
generating sufficiently high current-densities to provide a
large current in a current-matched dual-bandgap system.
4 Conclusion
Due to an improvement in light absorption, charge separation efficiency, high surface-areas for electrochemical reactions, and the ability to fabricate high-quality
single-crystalline materials with high carrier mobility,
semiconductor NWs are an attractive building block for
solar photovoltaic and photoelectrochemical energy conversion. We have focused on a combination of developing
large-scale NW arrays and single-NW devices to gain an
improved understanding of the fundamental properties
of these materials. In all cases, the end goal is to reduce
the cost and improve the energy conversion efficiency,
to enable large-scale adaptation of renewable energy.
Through the use of less raw material, with lower purity
and carrier lifetime requirements, combined with the
favorable optical and electrical properties of NWs, hold
great promise to enable a new generation of low-cost,
high efficiency solar energy conversion devices.
Acknowledgements N. P. Dasgupta acknowledges support from
the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy (EERE) Postdoctoral Research Awards under the
SunShot Solar Energy Technologies Program. This work was supported by the Director, Office of Science, Office of Basic Energy
Sciences, Materials Sciences and Engineering Division, of the U.S.
Department of Energy under Contract No. DE-AC02-05CH11231.
Neil P. Dasgupta and Peidong Yang, Front. Phys.
References
13
24.
M. D. Kelzenberg, D. B. Turner-Evans, B. M. Kayes, M. A.
Filler, M. C. Putnam, N. S. Lewis, and H. A. Atwater, Nano
Lett., 2008, 8(2): 710
1.
E. C. Garnett, M. L. Brongersma, Y. Cui, and M. D. McGehee, Annu. Rev. Mater. Res., 2011, 41(1): 269
25.
B. Z. Tian, X. Zheng, T. J. Kempa, Y. Fang, N. Yu, G. Yu,
J. Huang, and C. M. Lieber, Nature, 2007, 449(7164): 885
2.
A. I. Hochbaum and P. Yang, Chem. Rev., 2010, 110(1): 527
26.
3.
B. M. Kayes, H. A. Atwater, and N. S. Lewis, J. Appl. Phys.,
2005, 97(11): 114302
S. D. Oosterhout, M. M. Wienk, S. S. van Bavel, R. Thiedmann, L. Jan Anton Koster, J. Gilot, J. Loos, V. Schmidt,
and R. A. J. Janssen, Nat. Mater., 2009, 8(10): 818
27.
A. L. Briseno, T. W. Holcombe, A. I. Boukai, E. C. Garnett,
S. W. Shelton, J. J. M. Fréchet, and P. Yang, Nano Lett.,
2010, 10(1): 334
28.
J. A. Czaban, D. A. Thompson, and R. R. Lapierre, Nano
Lett., 2009, 9(1): 148
4.
L. Hu and G. Chen, Nano Lett., 2007, 7(11): 3249
5.
E. Garnett and P. Yang, Nano Lett., 2010, 10(3): 1082
6.
N. P. Dasgupta, S. Xu, H. J. Jung, A. Iancu, R. Fasching, R.
Sinclair, and F. B. Prinz, Adv. Funct. Mater., 2012, 22(17):
3650
7.
O. L. Muskens, J. G. Rivas, R. E. Algra, E. P. A. M. Bakkers,
and A. Lagendijk, Nano Lett., 2008, 8(9): 2638
29.
J. Tang, Z. Huo, S. Brittman, H. Gao, and P. Yang, Nat.
Nanotechnol., 2011, 6(9): 568
8.
L. Tsakalakos, J. Balch, J. Fronheiser, B. A. Korevaar,
O. Sulima, and J. Rand, Appl. Phys. Lett., 2007, 91(23):
233117
30.
L. Cao, P. Fan, A. P. Vasudev, J. S. White, Z. Yu, W. Cai,
J. A. Schuller, S. Fan, and M. L. Brongersma, Nano Lett.,
2010, 10(2): 439
9.
M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D.
Dunlop, Prog. Photovolt. Res. Appl., 2012, 20(5): 606
31.
10.
L. E. Greene, M. Law, D. H. Tan, M. Montano, J. Goldberger, G. Somorjai, and P. Yang, Nano Lett., 2005, 5(7):
1231
L. Cao, J. S. White, J. S. Park, J. A. Schuller, B. M.
Clemens, and M. L. Brongersma, Nat. Mater., 2009, 8(8):
643
32.
V. E. Ferry, L. A. Sweatlock, D. Pacifici, and H. A. Atwater,
Nano Lett., 2008, 8(12): 4391
33.
K. Nakayama, K. Tanabe, and H. A. Atwater, Appl. Phys.
Lett., 2008, 93(12): 121904
34.
H. A. Atwater and A. Polman, Nat. Mater., 2010, 9(3): 205
35.
S. Brittman, H. Gao, E. C. Garnett, and P. Yang, Nano
Lett., 2011, 11(12): 5189
36.
N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. USA,
2006, 103(43): 15729
37.
A. Listorti, J. Durrant, and J. Barber, Nat. Mater., 2009,
8(12): 929
38.
P. Yang, MRS Bull., 2012, 37(9): 806
39.
M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher,
Q. Mi, E. A. Santori, and N. S. Lewis, Chem. Rev., 2010,
110(11): 6446
40.
A. Fujishima and K. Honda, Nature, 1972, 238(5358): 37
41.
A. J. Nozik, Appl. Phys. Lett., 1976, 29(3): 150
42.
K. Ohashi, J. Mccann, and J. O. M. Bockris, Nature, 1977,
266(5603): 610
43.
A. Kudo, MRS Bull., 2011, 36(1): 32
44.
S. W. Boettcher, J. M. Spurgeon, M. C. Putnam, E. L. Warren, D. B. Turner-Evans, M. D. Kelzenberg, J. R. Maiolo,
H. A. Atwater, and N. S. Lewis, Science, 2010, 327(5962):
185
45.
S. W. Boettcher, E. L. Warren, M. C. Putnam, E. A. Santori, D. Turner-Evans, M. D. Kelzenberg, M. G. Walter, J.
R. McKone, B. S. Brunschwig, H. A. Atwater, and N. S.
Lewis, J. Am. Chem. Soc., 2011, 133(5): 1216
11.
L. E. Greene, B. D. Yuhas, M. Law, D. Zitoun, and P. Yang,
Inorg. Chem., 2006, 45(19): 7535
12.
M. Law, L. E. Greene, J. C. Johnson, R. Saykally, and P.
Yang, Nat. Mater., 2005, 4(6): 455
13.
A. C. Fisher, L. M. Peter, E. A. Ponomarev, A. B. Walker,
and K. G. U. Wijayantha, J. Phys. Chem. B, 2000, 104(5):
949
14.
M. Law, L. E. Greene, A. Radenovic, T. Kuykendall, J.
Liphardt, and P. Yang, J. Phys. Chem. B, 2006, 110(45):
22652
15.
L. E. Greene, M. Law, B. D. Yuhas, and P. Yang, J. Phys.
Chem. C, 2007, 111(50): 18451
16.
B. D. Yuhas and P. D. Yang, J. Am. Chem. Soc., 2009,
131(10): 3756
17.
M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D.
B. Turner-Evans, M. C. Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis, and H. A. Atwater, Nat.
Mater., 2010, 9(3): 239
18.
J. Zhu, C. M. Hsu, Z. Yu, S. Fan, and Y. Cui, Nano Lett.,
2010, 10(6): 1979
19.
J. Zhu, Z. Yu, G. F. Burkhard, C. M. Hsu, S. T. Connor, Y.
Xu, Q. Wang, M. McGehee, S. Fan, and Y. Cui, Nano Lett.,
2009, 9(1): 279
20.
E. Yablonovitch and G. D. Cody, IEEE Trans. Electron.
Dev., 1982, 29(2): 300
21.
M. G. Mauk, J. Miner. Met. Mater. Soc., 2003, 55(5): 38
46.
22.
A. Boukai, P. Haney, A. Katzenmeyer, G. M. Gallatin, A.
A. Talin, and P. Yang, Chem. Phys. Lett., 2011, 501(4–6):
153
A. Heller, E. Aharon-Shalom, W. A. Bonner, and B. Miller,
J. Am. Chem. Soc., 1982, 104(25): 6942
47.
Y. Hou, B. L. Abrams, P. C. K. Vesborg, M. E. Björketun,
K. Herbst, L. Bech, A. M. Setti, C. D. Damsgaard, T. Pedersen, O. Hansen, J. Rossmeisl, S. Dahl, J. K. Nørskov, and
I. Chorkendorff, Nat. Mater., 2011, 10(6): 434
23.
B. Tian, T. J. Kempa, and C. M. Lieber, Chem. Soc. Rev.,
2009, 38(1): 16
14
Neil P. Dasgupta and Peidong Yang, Front. Phys.
48.
B. Hinnemann, P. G. Moses, J. Bonde, K. P. Jørgensen, J.
H. Nielsen, S. Horch, I. Chorkendorff, and J. K. Nørskov, J.
Am. Chem. Soc., 2005, 127(15): 5308
54.
Y. W. Chen, J. D. Prange, S. Dühnen, Y. Park, M. Gunji,
C. E. D. Chidsey, and P. C. McIntyre, Nat. Mater., 2011,
10(7): 539
49.
M. Tomkiewicz and J. M. Woodall, Science, 1977, 196(4293):
990
55.
Y. J. Hwang, A. Boukai, and P. D. Yang, Nano Lett., 2009,
9(1): 410
50.
J. Sun, C. Liu, and P. Yang, J. Am. Chem. Soc., 2011,
133(48): 19306
56.
T. Kuykendall, P. Ulrich, S. Aloni, and P. Yang, Nat. Mater.,
2007, 6(12): 951
51.
C. Liu, J. Sun, J. Tang, and P. Yang, Nano Lett., 2012,
12(10): 5407
57.
52.
Y. J. Hwang, C. Hahn, B. Liu, and P. Yang, ACS Nano,
2012, 6(6): 5060
Y. J. Hwang, C. H. Wu, C. Hahn, H. E. Jeong, and P. Yang,
Nano Lett., 2012, 12(3): 1678
58.
F. Le Formal, N. Tétreault, M. Cornuz, T. Moehl, M.
Grätzel, and K. Sivula, Chem. Sci., 2011, 2(4): 737
C. Liu, Y. J. Hwang, H. E. Jeong, and P. Yang, Nano Lett.,
2011, 11(9): 3755
59.
P. Yang and J. M. Tarascon, Nat. Mater., 2012, 11(7): 560
53.